Humans in the Loop of Localization Processes

Abstract

AbstractThe Language Grid is a service-oriented infrastructure for language services. In the Language Grid, machine translation services play important roles in supporting multilingual activities for communities. Although the effectiveness of using machine translation services for multilingual communication has been shown in previous reports, the gap between human translators and machine translators remains huge especially in the domain of localization processes that require high translation quality. In this chapter, we aim at improving localization processes by introducing humans into the loop to utilize machine translation services. We try to compare several different types of localization processes (i.e., absolute machine translation processes, absolute human translation processes and processes by human and machine translation services) in the dimensions of translation quality and translation cost. The experiment results show that monolinguals can help improve the translation quality of machine translators with the aid of community dictionary services, and that collaboration of human and machine translation services make it possible to reduce the cost compared with absolute human translations.KeywordsLocalization ProcessMachine TranslationComposite ServiceProcess InstanceExecution CostThese keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.